Thanks Jim Mulder. Perhaps you can provide guidance if I explain what I thought I was doing.
At times there's a bit of code which "runs like a dog with no legs". There will usually be several options to deal with it, but to actually land on which is "better" you need to know how they perform against each other. Rather than tossing five different versions of the program onto the machine, collating usage, and not being able to compare with the "overhead" of the rest of the program, I was hoping to get "a starting point" of the CPU used, an "ending point" of the CPU used, thus being able to arrive at "the CPU used by this bit of code" (repeated lots of times). I'm not concerned with IO of any type, including DBs, I am concerned with explicit CALL and implicit (COBOL run-time routines). Nothing fancy, although it can include LE callable services. I'm not interested in System Services. (In summary, I'm not interested in things (at that particular stage) that I can have no influence on, if I have to read a KSDS there's only one way to do it in COBOL, so I "abstract it out" of the loop). Using it so far does (apparently) give fair indicative results. "Yes, that's significantly better" or "yes, that's faster, but not enough to chose it in this case over something "less exotic". If there's some way to make it more "accurate" for what I want, I'd be pleased. I'm "applications", so when you say "despatch", I think "mail room" or "loading bay". As plain as possible is fine with me. Perhaps-related question. "interruptible" instructions (like MVCL, I was told), are they "doing me in"? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
